Simple Entity Component System implementation with minimal robotics application examples.
This project is mainly for learning and exploration purposes. For physics (and collision) system, lazyECS utilizes ReactPhysics3D (https://github.com/DanielChappuis/reactphysics3d).
Important Note: If the update rate of the application is faster than the Render rate (which is currently 50Hz), then the PhysicsSystem->Update() function better be called in the main application loop. (so that latest forces/teleportations can be applied immediately and interpolation factor introduces less error)
If the update rate of the application is slower then the Render rate, PhysicsSystem->Update() function better be called outside the main application loop, which is running faster (at the rate of Render System), so that interpolation factor has a resolution that fits better to the render rate.
Minimal lazyECS application with both physics and render only actors that can be moved around kinematically or dynamically.
Naive implementation of the potential field algorithm where Green capsule represent the goal and Red capsules represent the obstacles. Obstacles have repulsive and goal has attractive force field attached to them, which are calculated based on Normal Distribution, where mean of the distribution is the center of the goal/obstacle location and sigma defines the propagation of the field around the object. When/how this naive implementation suffer from being stuck in local minima is quite dependent on the layout of obstacles and the sigma of the distribution.
2D Grid based implementation of path finding by wave propagation. Compared to the potential field, implementation of this approach is more "discrete" in the sense that, the environment is divided into finite number of cells, each cell is either "occupied" or "free". There is a cost value associated with each free cell based on it's relative distance to the goal position, and the ego actor is trying to move down the cost gradient in the fastest way possible, by choosing the free cell with the lowest "cost" (distance) value in it's 4-directional neighborhood. Occupied cells omitted from the search.
